Abstract: Aspects of the present disclosure provide a method for training a predictor that predicts performance of a plurality of machine learning (ML) models on platforms. For example, the method can include converting each of the ML models into a plurality of instructions or the instructions and a plurality of intermediate representations (IRs). The method can also include simulating execution of the instructions corresponding to each of the ML models on a platform and generating instruction performance reports. Each of the instruction performance reports can be associated with performance of the instructions corresponding to one of the ML models that are executed on the platform. The method can also include training the predictor with the instructions or the IRs as learning features and the instruction performance reports as learning labels, compiling the predictor into a library file, and storing the library file in a storage device.

Abstract: The present disclosure relates to an embedded flash memory cell having a common source oxide layer with a substantially flat top surface, disposed between a common source region and a common erase gate, and a method of formation. In some embodiments, the embedded flash memory cell has a semiconductor substrate with a common source region separated from a first drain region by a first channel region and separated from a second drain region by a second channel region. A high-quality common source oxide layer is formed by an in-situ steam generation (ISSG) process at a location overlying the common source region. First and second floating gate are disposed over the first and second channel regions on opposing sides of a common erase gate having a substantially flat bottom surface abutting a substantially flat top surface of the common source oxide layer.

Abstract: The present disclosure relates to an embedded flash memory cell having a common source oxide layer with a substantially flat top surface, disposed between a common source region and a common erase gate, and a method of formation. In some embodiments, the embedded flash memory cell has a semiconductor substrate with a common source region separated from a first drain region by a first channel region and separated from a second drain region by a second channel region. A high-quality common source oxide layer is formed by an in-situ steam generation (ISSG) process at a location overlying the common source region. First and second floating gate are disposed over the first and second channel regions on opposing sides of a common erase gate having a substantially flat bottom surface abutting a substantially flat top surface of the common source oxide layer.